arXiv subjects

Neha Karanjkar

Publications and source records attributed to Neha Karanjkar.

5 recordsLinked to original sources

arXiv2026

On the Potential of Graph Neural Networks as Metamodels for Supply Chain Optimization: Dataset, Architectures, and Directions

Graph Neural Networks (GNNs) have emerged as a powerful, differentiable class of learning models for graph-structured systems. Their ability to generalize across topologies opens the prospect of a surrogate for combined structural and parametric optimization, which classical metamodels cannot offer. Supply chains are a natural target, yet the use of GNN surrogates for supply chain problems is largely unexplored. This paper lays the foundation, presents initial steps, and discusses key research directions. As a foundation, we formulate the problem and create a large public training dataset of programmatically generated supply chain graphs with input parameters and steady-state performance metrics obtained using our SupplyNetPy simulation library. As initial steps, we explore GNN architectures that work well as surrogates for node- and network-level predictions, and analyze their accuracy-compute trade-off against simulation. Most importantly, we outline the exciting directions this opens, namely gradient-based optimization over topology, fast design-space exploration, and sensitivity analysis.

cs.LG↗

arXiv2026

SupplyNetPy: An Open-Source Python Library for High-Fidelity Modeling and Simulation of Arbitrary Supply Chain and Inventory Networks

This paper introduces SupplyNetPy, an open-source, well-documented Python library for modeling and discrete-event simulation of supply chain networks with arbitrary multi-echelon structures. It supports multiple replenishment policies, perishable inventory, node disruptions, and stochastic demand and lead times. All components are extensible via inheritance. Users describe a supply chain as a graph with node and link attributes, while the library handles simulation, providing logs and extensive node and network level performance reports. This paper presents the motivation, design, key features, and architecture of SupplyNetPy, along with detailed validation results (against analytical benchmarks, a commercial tool, and a published case study). A key motivation behind SupplyNetPy's development is programmatic generation and simulation of complex models, enabling design-space exploration, what-if analysis, training data generation, and supply chain digital twins.

cs.AI↗

arXiv2026

On Integrating Resilience and Human Oversight into LLM-Assisted Modeling Workflows for Digital Twins

LLM-assisted modeling holds the potential to rapidly build executable Digital Twins of complex systems from only coarse descriptions and sensor data. However, resilience to LLM hallucination, human oversight, and real-time model adaptability remain challenging and often mutually conflicting requirements. We present three critical design principles for integrating resilience and oversight into such workflows, derived from insights gained through our work on FactoryFlow - an open-source LLM-assisted framework for building simulation-based Digital Twins of manufacturing systems. First, orthogonalize structural modeling and parameter fitting. Structural descriptions (components, interconnections) are LLM-translated from coarse natural language to an intermediate representation (IR) with human visualization and validation, which is algorithmically converted to the final model. Parameter inference, in contrast, operates continuously on sensor data streams with expert-tunable controls. Second, restrict the model IR to interconnections of parameterized, pre-validated library components rather than monolithic simulation code, enabling interpretability and error-resilience. Third, and most important, is to use a density-preserving IR. When IR descriptions expand dramatically from compact inputs hallucination errors accumulate proportionally. We present the case for Python as a density-preserving IR : loops express regularity compactly, classes capture hierarchy and composition, and the result remains highly readable while exploiting LLMs strong code generation capabilities. A key contribution is detailed characterization of LLM-induced errors across model descriptions of varying detail and complexity, revealing how IR choice critically impacts error rates. These insights provide actionable guidance for building resilient and transparent LLM-assisted simulation automation workflows.

eess.SY↗

arXiv2022

A Python-based Mixed Discrete-Continuous Simulation Framework for Digital Twins

The use of Digital Twins is set to transform the manufacturing sector by aiding monitoring and real-time decision making. For several applications in this sector, the system to be modeled consists of a mix of discrete-event and continuous processes interacting with each other. Building simulation-based Digital Twins of such systems necessitates an open, flexible simulation framework which can support easy modeling and fast simulation of both continuous and discrete-event components, and their interactions. In this paper, we present an outline and key design aspects of a Python-based framework for performing mixed discrete-continuous simulations. The continuous processes in the system are assumed to be loosely coupled to other components via pre-defined events. For example, a continuous state variable crossing a threshold may trigger an external event. Similarly, external events may lead to a sudden change in the trajectory, state value or boundary conditions in a continuous process. We first present a systematic events-based interface using which such interactions can be modeled and simulated. We then discuss implementation details of the framework along with a detailed example. In our implementation, the advancement of time is controlled and performed using the event-stepped engine of SimPy (a popular discrete-event simulation library in Python). The continuous processes are modelled using existing frameworks with a Python wrapper providing the events interface. We discuss possible improvements to the time advancement scheme, a roadmap and use cases for the framework.

eess.SY↗

arXiv2018

On Continuous-space Embedding of Discrete-parameter Queueing Systems

Motivated by the problem of discrete-parameter simulation optimization (DPSO) of queueing systems, we consider the problem of embedding the discrete parameter space into a continuous one so that descent-based continuous-space methods could be directly applied for efficient optimization. We show that a randomization of the simulation model itself can be used to achieve such an embedding when the objective function is a long-run average measure. Unlike spatial interpolation, the computational cost of this embedding is independent of the number of parameters in the system, making the approach ideally suited to high-dimensional problems. We describe in detail the application of this technique to discrete-time queues for embedding queue capacities, number of servers and server-delay parameters into continuous space and empirically show that the technique can produce smooth interpolations of the objective function. Through an optimization case-study of a queueing network with $10^7$ design points, we demonstrate that existing continuous optimizers can be effectively applied over such an embedding to find good solutions.

cs.PF↗